Characterization of Hyperinsulinism in Infancy Assessed with PET and 18F-Fluoro-L-DOPA Maria-Joa˜o Ribeiro, MD, PhD1; Pascale De Lonlay, MD, PhD2; Thierry Delzescaux, PhD1; Nathalie Boddaert, MD3; Francis Jaubert, MD4; Sandrine Bourgeois1; Fre´de´ric Dolle´, PhD1; Claire Nihoul-Fe´ke´te´, MD5; Andre´ Syrota, MD, PhD1; and Francis Brunelle, MD3 1Service Hospitalier Fre´de´ric Joliot, De´partement de Recherche Me´dicale, Direction des Sciences du Vivant, Commissariat a` l’Energie Atomique, Orsay, France; 2De´partement de Me´tabolisme et Pe´diatrie, Hoˆpital Necker-Enfants Malades, Paris, France; 3Service de Radiologie Pe´diatrique, Hoˆpital Necker-Enfants Malades, Paris, France; 4Laboratoire de Anatomopathologie, Hoˆpital Necker-Enfants Malades, Paris, France; and 5De´partement de Chirurgie Infantile, Hoˆpital Necker-Enfants Malades, Paris, France Hyperinsulinism (HI) of infancy is a neuroendocrine disease Hyperinsulinism (HI) is the most important cause of secondary to either focal adenomatous hyperplasia or a diffuse recurrent hypoglycemia in infancy. The hypersecretion of abnormality of insulin secretion of the pancreas. HI with focal insulin induces profound hypoglycemias that require ag- lesions can revert by selective surgical resection in contrast to gressive treatment to prevent the high risk of neurologic the diffuse form, which requires subtotal pancreatectomy when resistant to medical treatment. Neuroendocrine diseases are a complications (1,2). HI can be due to 2 different histopatho- heterogeneous group of entities with the ability to take up amine logic types of lesions, a focal or a diffuse form (3,4), based precursors and to convert them into biogenic amines. There- on different molecular entities despite an indistinguishable fore, the aim of this study was (a) to evaluate the use of PET with clinical pattern (5–9). In focal HI, which represents about 18F-fluoro-L-dihydroxyphenylalanine (18F-fluoro-L-DOPA) and (b) 40% of all cases (10), the pathologic pancreatic ␤-cells are to distinguish between focal and diffuse HI. Methods: Fifteen gathered in a focal adenoma, usually 2.5–7.5 mm in diam- patients (11 boys, 4 girls) with neonatal HI were enrolled in this eter. Conversely, diffuse HI corresponds to an abnormal study. All patients fasted for at least 6 h before the PET exam- insulin secretion of the whole pancreas with disseminated ination and their medication was discontinued for at least 72 h. ␤ The examination was performed under light sedation (pentobar- -cells showing enlarged abnormal nuclei (11). Finally, bital associated with or without chloral). The dynamic acquisi- about 10% of HI cases are clinically atypical and cannot be tion started 45–65 min after the injection of 18F-fluoro-L-DOPA classified, having an unknown molecular basis and his- (4.0 MBq/kg weight). Four or 6 scans of 5 min each (2 or 3 steps topathologic form (12). according to the height of the patient) were acquired from the Control of HI is attempted through medical treatment neck to the upper legs. Results: An abnormal focal pancreatic with diazoxide, nifedipine, or octreotide (13–15), but pan- uptake of 18F-fluoro-L-DOPA was observed in 5 patients, createctomy is the only issue for patients who are resistant whereas a diffuse uptake of the radiotracer was observed in the pancreatic area of the other patients. All patients with focal to these treatments (10,16). Therefore, the differential diag- radiotracer uptake and also 4 of 10 patients with pancreatic nosis between the 2 forms becomes of major importance diffuse radiotracer accumulation, unresponsive to medical treat- since their surgical treatment and the outcome differ con- ment, underwent surgery. The histopathologic results confirmed siderably. Focal HI is totally cured by selective resection of the PET findings—that is, focal versus diffuse HI. Conclusion: the adenoma, whereas diffuse forms of HI require a subtotal The results of this study suggest that 18F-fluoro-L-DOPA could pancreatectomy, with severe iatrogenic diabetes as a con- be an accurate noninvasive technique to distinguish between sequence (17,18). focal and diffuse forms of HI. The localization of insulin hypersecretion before surgery 18 Key Words: hyperinsulinism of infancy; PET; F-fluoro-L-DOPA is possible only through pancreatic venous catheterization J Nucl Med 2005; 46:560–566 (PVS), allowing a pancreatic map of insulin concentrations, with an eventual additional pancreatic arterial calcium stim- ulation (19–21). PVS is an invasive method, which is tech- nically difficult to perform and requires general anesthesia. The concentrations of plasmatic glucose must be maintained between 2 and 3 mmol/L before and during PVS. Moreover, Received Sep. 9, 2004; revision accepted Dec. 15, 2004. For correspondence or reprints contact: Maria-Joa˜ o Ribeiro, MD, PhD, all medical treatments must be stopped 5 d before the study. Service Hospitalier Fre´de´ ric Joliot, DRM/DSV, CEA, 4, place du Ge´ne´ ral Therefore, it is of major interest to find a less invasive way Leclerc, F-91406 Orsay, France. E-mail: [email protected] to differentiate between focal and diffuse HI. This method 560 THE JOURNAL OF NUCLEAR MEDICINE • Vol. 46 • No. 4 • April 2005 should precisely localize the pathologic area of focal HI to MATERIALS AND METHODS guide the surgeon. Patients L-Dihydroxyphenylalanine (L-DOPA) is a precursor of Fifteen patients (11 boys, 4 girls; age range, 1–14 mo; mean catecholamines that is converted to dopamine by the aro- age Ϯ SD, 5.7 Ϯ 3.8 mo) with neonatal HI were studied with matic amino acid decarboxylase (AADC). In addition to its 18F-fluoro-L-DOPA and PET (Table 1). In all cases, HI diagnosis role as a precursor of noradrenaline and adrenaline, dopa- was based on persistent hypoglycemia with low plasma ketone mine is a transmitter substance in the central and peripheral bodies and free fatty acids together with measurable circulating nervous system. The capacity to take up and decarboxylate insulin levels at the time of hypoglycemia. amine precursors such as L-DOPA and 5-hydroxytrypto- All patients fasted for at least 6 h before the PET study and their phan and to store their biogenic amine (dopamine and medications were stopped for at least 72 h. Two patients had 2 scans. One patient had a PET study before and after treatment with serotonin) is characteristic of neuroendocrine cells. an inhibitor of AADC (L-␣-hydrazino-␣-methyl-␤-(3,4-dihydroxy- Pancreatic cells contain markers usually associated with phenyl)propionic acid or carbidopa). Another patient had a PET neuroendocrine cells, such as tyrosine hydroxylase, dopa- study 72 h after drug withdrawal and another PET study after mine, neuronal dopamine transporter, vesicular dopamine administration of octreotide and diazoxide. During all PET studies, transporter, and monoamine oxidases A and B (22–24). normoglycemia was maintained by glucose infusion, which was Pancreatic islets have been shown to take up L-DOPA and carefully adjusted according to frequent blood glucose monitoring. convert it to dopamine through the AADC (25–27). Maximal glucose infusion rates between 6.4 and 13.2 mg/kg/min PET performed with 18F-fluoro-L-dihydroxyphenylala- were needed. PET acquisition was performed under light sedation nine (18F-fluoro-L-DOPA) has been extensively used to (pentobarbital associated with or without chloral). study the central dopaminergic system. Nevertheless, sev- The 5 patients for whom 18F-fluoro-L-DOPA uptake results eral recent studies have demonstrated the usefulness of this strongly suggested focal HI and the patients with diffuse HI ϭ radiotracer to detect neuroendocrine tumors such as pheo- resistant to medical treatment (n 4/10) underwent surgery. chromocytomas, thyroid medullar carcinomas, or gastroin- Pancreatic tissue obtained from surgical resections was fixed in formalin and embedded in paraffin, and serial sections were stud- testinal carcinoid tumors that usually contain secretory ied by immunohistochemistry after a water bath antigen retrieval granules and have the ability to produce biogenic amines step. The primary antibodies used were antiproinsulin (1/400 (28,29). mouse monoclonal antibody, 1G4; Novocastra), antichromogranin The aim of this work was to evaluate the use of whole- A (1/200 mouse monoclonal antibody DAK-A3; DAKO), antisyn- body PET with 18F-fluoro-L-DOPA to detect the hyperfunc- aptophysin (1/50 rabbit polyclonal antibody A0010; DAKO), and tional pancreatic islet tissue and to test its ability to differ- anti-DOPA decarboxylase or anti-AADC (1/100 rabbit polyclonal entiate between focal and diffuse HI. antibody; Chemicon International). TABLE 1 Clinical Profile of 15 Patients with HI Birth Age (mo) Weight at Patient weight Gestational At At PET PET scan Response to no. Sex (g) age (wk) diagnosis scan (g) medication Surgery 1 M 3,470 36 Neonatal 5 7,500 DzxϪ, OctrϪ Yes 2 M 3,760 40 Neonatal 3 5,800 DzxϪ, OctrϪ Yes 3 F 3,310 38 Neonatal 2 5,000 DzxϪ, OctrϪ Yes 4 M 3,871 38 Neonatal 4 9,160 DzxϪ, OctrϪ Yes 5 M 4,050 37 Neonatal 7 10,870 DzxϪ, Octrϩ Yes 6 M 4,450 38 Neonatal 4 8,190 DzxϪ, OctrϪ Yes 7 M 3,690 36 Neonatal 5 8,600 DzxϪ, Octrϩ No 8 M 3,900 39 Neonatal 3 and 12 6,800 DzxϪ, Octrϩ No 10,680 9 F 2,950 37 2.5 13 11,000 Dzxϩ No 10 F 4,010 37 Neonatal 8 7,300 DzxϪ, Octrϩ No 11 F 3,480 37 Neonatal 5 6,070 DzxϪ, OctrϪ Yes 12 M 5,000 38 Neonatal 14 and 17 11,500 DzxϪ, OctrϪ Yes 12,200 13 M 3,700 38 Neonatal 2 6,670 DzxϪ, Octrϩ No 14 M 3,800 38 Neonatal 9 9,800 DzxϪ, Octrϩ No 15 M 3,580 36 Neonatal 2 6,000 DzxϪ, OctrϪ Yes Dzx ϭ diazoxide; Octr ϭ octreotide. HYPERINSULINISM OF INFANCY AND PET • Ribeiro et al.